Print

[gnh1201]

Hi.

I modified "jppf-gui.properties" file in admin-UI as below.

Code: [Select]

driver1.jppf.server.host = myhost.remote

Or

Code: [Select]

driver1.jppf.server.host = xxx.xxx.xxx.xxx

The server is working well, too.

Code: [Select]

driver process id: 8094, uuid: 25390EE9-0E77-4BD9-93CC-13BDA82C69D2
ClientClassServer initialized
NodeClassServer initialized
ClientJobServer initialized
NodeJobServer initialized
management initialized and listening on port 11111
secure management initialized and listening on port 11443
Acceptor initialized
-  accepting plain connections on port 11111
- accepting secure connections on port 11443
JPPF Driver initialization complete

I turned off all port-based firewalls in all routes (Windows Desktop and my Linux server). I checked with Telnet and there is no problem with 11111, 11443 ports.

However, JPPF admin-UI searches only the localhost server and does not even attempt to access the remote address.

If the server is on localhost, it is scanned well.

I attached the log, so I'd like to know what the problem is.

[lolo]

Hello,

Looking at the logs you provided, it looks like your admin console is configured to discover the driver through the built-in UDP multicast dsicovery mechanism.
I would suggest that you disable it (jppf.discovery.enabled = false) et configure the connection manually in the configuration file instead, as in this example:

Code: [Select]

jppf.discovery.enabled = false
jppf.drivers = driver1
driver1.jppf.server.host = <ip_address_or_host_name>
driver1.jppf.server.port = 11111
Could you please try that and let us know if this works?

Sincerely,
-Laurent

[gnh1201]

It works! Thanks

Code: [Select]

jppf.discovery.enabled = false


[caffeined]

Hello Laurent, jnh1201,

@Laurent - thanks for that solution! Turning off the UDP discovery also did the trick for me.  Began working right away!

I wanted to also add some interesting details from my experience.

I have a small raspberry pi cluster cluster and was running the driver on one of them.  From my laptop i was attempting to use the admin gui and failing to connect.  Since the laptop i am using is a work machine it has firewall set up and i assumed it was a firewall issue.  As a test of this i attempted to have a node connect to the driver from my laptop.  To my surprise it worked fine.

I looked at the config for the node, expecting to find that its default is to *not* do UDP discovery-  but, no, it also uses that.

So I guess that maybe the node is 'smart' enough to go ahead with the explicit host/port if they're in the config file (jppf-node.properties), but the admin gui isn't?

I'd be curious if anyone can explain the difference in behavior - I'd just like to understand.  It's working now, so I'm going to go and play with my raspberry pi cluster, but if anyone has a comment i'd be interested....

(and thanks again to Laurent Cohen for all his work in developing JPPF!)